Elsevier

Chemical Engineering Science

Volume 138, 22 December 2015, Pages 814-826
Chemical Engineering Science

Customization of an optical probe device and validation of a signal processing procedure to study gas–liquid–solid flows. Application to a three-phase internal-loop gas-lift Bioreactor

https://doi.org/10.1016/j.ces.2015.08.046Get rights and content

Highlights

  • Customization of optical device for application in three-phase systems.

  • Application and validation of new technique at lower solids content.

  • Solids decrease gas hold-up in riser and increase system complexity.

  • Validation of results using additional techniques.

Abstract

The study of local hydrodynamic properties of three-phase bioreactors in biotechnology processes is of great importance, mainly because of the complex interaction between bioreactor and microorganisms. However, classical techniques used for measuring local hydrodynamic properties such as single needle probes are mainly limited to two-phase flows. In this work it was developed and validated a new system, based on the customization of an optical probe initially designed in LEGI. The necessity of a new system was due to the agglomeration of the solid-phase (spent grains which are used as the micro-organisms carrier for the targeted application) around the optical tip, which influences the measurements. This new system allows for the measurement of the main local gas-phase properties in a complex gas–liquid–solid mixture. The new system was first validated for air–water system in an internal loop gas-lift reactor and then applied to a spent grains–air–water mixture at low solid load in an internal gas lift reactor. In addition, experiments using complementary techniques (as high speed camera and PIV) were performed that allowed for the validation of the new system and the explanation of possible physical mechanisms that are underlying on this multiphase system. The system developed has the potential for improvement and use in several biotechnology applications.

Introduction

The study of gas-lift reactors hydrodynamics has been increasing over the last decades, especially due to the several biotechnological applications of these reactors. These applications are called (g–l–s) systems as they are composed by three phases: gas (g), liquid (l) and solid (s). In some cases the inert solid phase can act as the carrier for immobilization of living biomass.

More specifically, in the long term, we are interested in continuous high cell density fermentation systems using immobilized cell technology: the lack of knowledge about these bioreactor hydrodynamics hampered their industrial development.

To better understand the local hydrodynamics of such a complex system, in particular the effect of the solid and biomass on the gas–liquid (g–l) mixture, the characterization of each phase requires the use of different and refined measurement techniques. The techniques for studying the local gas-phase properties can be divided into two categories: non-invasive and invasive. Among non-invasive techniques we can list: particle image velocimetry (PIV), particle image tracking (PIT) and visualization techniques. On the other hand, invasive techniques for studying local gas-phase properties use mainly phase detection probes (Boyer et al., 2002, Cartellier, 1992). Most of the invasive techniques, such as phase detection probes, are cheaper and easier to apply in industrial processes than non-invasive ones. The reasons for this are: (1) the typical turbulent flow regimes found in industrial reactors; (2) opaque reactors and (3) the ability to continuously collect data online (Boyer et al., 2002). Sensors are based on temperature, capacitance, optical and conductivity properties of the system. Frequently, they are applied on two-phase flow(Jhawar and Prakash, 2007; Olivieri et al., 2007) and more recently to three-phase flow systems (Mena et al., 2008; Hooshyar et al., 2010; Wu et al., 2008), usually at low solids holdup (Boyer et al., 2002).

The present work focuses on the customization of an optical probe device to study gas–liquid–solid system. It is indeed a prerequisite before studying the complete bioreactor behaviour with biomass. More specifically, given the industrial applications targeted, the solid considered in this paper is spent grain whose properties are different from the solids used in previous work (Mena et al., 2008). Therefore specific developments were required. The experiment was performed in an internal loop gas lift reactor (iGLR) which also corresponds to the targeted industrial applications.

In the present manuscript, several measurements techniques were used. The liquid and solid phase hydrodynamic behaviour was characterized using PIV and a visualization technique respectively. In the frame of the gas phase characterization, two techniques were used: photographic method (high speed camera – denoted HSC in this paper) and optical probe (denoted OP). This choice was driven by the need for complementary measurement methods to assure the correctness of the optical probe data.

The main limitations of HSC are the inability to use it at high gas flows and at high solid loads. The reactor wall transparency, out of focus, illumination and image analysis are among other known limitations (Ferreira et al., 2012). The reactor wall, focus and illumination can be solved/reduced experimentally (Mena et al., 2008, Mena et al., 2005b). Concerning image processing, automatic image analysis is not an easy task mainly because of the identification of single and overlapping bubbles. Usually this is done by manual or semi-manual methods (Ferreira et al., 2012). Recently an automatic method has been developed, which allows the identification between a single and a group (≥2) of bubbles, however this is also limited under experimental conditions cited above. The classification is based on a series of probabilities and it also allows the characterization of the system׳s complexity (Ferreira et al., 2012). When bubbles are classified, the commercial programs available are then able to determine bubble properties such as: the projected area, equivalent diameter, the Feret diameters distribution, elongation and sphericity (Ferreira et al., 2012).

Among other measurement tools, over the last two decades optical probes have been developed and successfully applied in bubble column reactors (Cartellier, 1998, Hamad and He, 2010, Hong et al., 2004, Shen et al., 2008). Their operation is relatively simple: it is based on the change of the refractive index of the medium in which the sensitive tip is located. Depending on this refractive index, more or less light intensity is sent back to a detector: this allows the measurement of the phase indicator function. From this indicator function, it is possible to measure gas-phase concentrations (holdup). Under some conditions, it is also possible to measure bubble velocity and then, bubble size distributions, mean interfacial area and mean Sauter diameter. To measure parameters such as bubble velocities, diameter and interfacial area double (or multi) tip probes are often used (Boyer et al., 2002, Chaumat et al., 2005), however their major drawback is the low spatial resolution. Cartellier (1990) developed a single tip probe which allows an accurate description of the flow. This type of probes was used in our work and is able to measure void fraction and bubble velocity (Cartellier, 1990). Moreover, due to their small size the spatial resolution is very high. The optical probe principle as well as a complete description of the LEGI optical probes signal processing methods (treatment for gaseous velocity – TGV) and main measurement errors can be found in the literature (Cartellier, 1998, Cartellier, 1992, Mena et al., 2008). Let us just stress that the probe manufacturing allows to control the diameter and shape of the sensitive tip (which is an almost perfect cone). That explains their high spatial resolution (smaller bubble detected, better interface piercing). Because of this controlled conical shape, Cartellier (1990) showed that the interface piercing time could be related to the bubble velocity through a calibration curve. However, that does not completely supress some uncertainties on the data. Using one of the probes manufactured in LEGI, Vejrazka et al. (2010) found that gas holdup measurement errors obtained when this OP is used are still due to the intrusive character of these probes. They studied the interaction between an optical probe and a bubble: where Optical probe displacement and bubble deformation (leading to imperfect tip de-wetting, bubble deceleration or deflection) were found to be the causes for the measurement errors displayed, especially when chords distribution is calculated (Vejražka et al., 2010). However it is believed that the ability to evaluate gas holdup is not so roughly affected.

Initially the long term objective of this work was to study the local influence of spent grains carrying immobilized cells through the hydrodynamic properties of the three-phase iGLR. The cause of this interest is the recent developments on continuous high cell density fermentation systems using immobilized cell technology and the lack of knowledge associated with the bioreactor hydrodynamic aspects.

During the course of this work some difficulties related with measurement techniques arose, especially the use of OP in a three-phase g–l–s system whenever spent grains are used. Therefore the first step was to study the local hydrodynamics of a three-phase iGLR using spent grains as solid-phase. This included the development and calibration of a new method to measure gas-phase properties using an OP in a system where spent grains were present as well as the evaluation of the effect of low solid concentration in iGLR local hydrodynamics at low gas flow rates (UGr≤1 cm/s). This is the subject of the present paper.

Section snippets

Experimental apparatus

The iGLR used in this work is of the concentric draught tube type with an enlarged top section for degassing and a total working volume of 6 L (Fig. 1). The angle between the conical sector and the main body was 51°. Gas was injected through a distributor (1-cm diameter) with five needles, each of 0.2 mm in diameter and placed 1.7 cm below the annulus of the riser. The water level in the reactor was kept constant. Temperature and pressure were ambient (21 °C and 1 atm). The desired gas flow was

Optical probe application in gls systems with spent grains

The solids used in this work (spent grains) have the ability, when in the presence of water, to form aggregates especially in the sections of the reactor (conical part) where the liquid velocity is lower (dead zones): this behaviour results from their hydrophobic surface and rugosity. These particles and bubbles have similar sizes (1 mm to 5 mm) so the tip can pierce one or more particles that are flowing upwards in the riser. The pierced particles will then interact with the moving particles and

Conclusions

A new system based on the customization of an existing optical probe (equipped with a cleaning system using a periodic injection) and a new signal processing algorithm was developed in order to perform simultaneous measurement of the gas holdup, bubble size and velocity in a gas–liquid–solid airlift (with spent grains as solids) at various solids concentrations. The solids specificity was its tendency to create agglomerates around the probe tip. This measurement tool was then assessed through

Nomenclature

    Symbols

    A

    Cross section area (cm2)

    C

    Distribution factor (dimensionless)

    chB

    Bubble Chord (mm)

    d12

    Distance between two riser points (mm)

    deq

    Equivalent diameter (mm)

    H1H2

    Difference of water level in inverted manometer (mm)

    Q

    Gas flow rate (mL/min)

    r

    Radial position of the probe

    Re

    Reynold number (dimensionless)

    tA

    Time when the signal from OP is above the noise before tC (s)

    tB

    Time when the signal from OP starts declining (s)

    tC

    Time when the signal from OP starts rising (s)

    tD

    Time when the signal from OP reaches maximum

Acknowledgements

The authors gratefully acknowledge the financial support from FCT (Fundação para a Ciência e Tecnologia, SFRH/BD/37082/2007 and SFRH/BPD/45637/2008).

References (37)

Cited by (7)

  • Characterisation of transverse turbulent motion in quasi-two-dimensional aerated flow: Application of four-point air-water flow measurements in hydraulic jump

    2019, Experimental Thermal and Fluid Science
    Citation Excerpt :

    The development of bubbly flow imaging techniques also allowed for a depiction of two-dimensional velocity field in a longitudinal-vertical visualisation plane, although the bubble diffusion and velocity distribution may be distorted by the lateral boundary layer next to the observation window, and no information in the transverse direction is available [2,3,17,25]. Measurement of air-water flow turbulence properties in the spanwise direction may involve multi-point measurements, although application in high-speed open channel flows was rare [21,26]. For the particular case of highly-aerated hydraulic jumps, measure of transverse flow structures was investigated in terms of the integral turbulent length and time scales based upon correlation analyses.

  • Investigating bubble dynamics in a bubble column containing shear thinning liquid using a dual-tip probe

    2018, Experimental Thermal and Fluid Science
    Citation Excerpt :

    Although, the bubble rise velocity can be determined using a single-tip probe, a dual-tip probe can do the same with more accuracy. Measuring the bubble rise velocity by a single-tip probe contains two stages, (i) selecting two thresholds in the voltage signal (at rise of the voltage when the bubble is piercing the tip) and (ii) evaluating the rise time from rise of the voltage considering the chosen thresholds and estimating the bubble rise velocity by multiplying the rise time by a coefficient of proportionality obtained from the probe calibration [19,26]. Assessing the rise time from the voltage signal is affected by the level of thresholds as well as the liquid film drainage time of the probe tip [28–30].

  • Effects of operational models (batch, continuous and CFID modes) on the performance of a single A<inf>2</inf>O airlift bioreactor for treatment of milk processing wastewater

    2017, Chemical Engineering Research and Design
    Citation Excerpt :

    The airlift reactor (ALR) is an interesting reactor which enhances the contact efficiency between gas–liquid or gas–liquid–solid phases through fluid circulation (Baz-Rodríguez et al., 2014; Mota et al., 2015).

  • Simultaneous measurements of two phases using an optical probe

    2019, Experimental and Computational Multiphase Flow
  • Modeling an airlift reactor for the growing of microalgae

    2018, Open Chemical Engineering Journal
View all citing articles on Scopus
View full text